dsgnPrep: Pre-process the Englert and Kieser (2013) optimal adaptive...

In arsenionhacolo/InferenceBEAGSD: Inference for Binary Endpoint Adaptive Group-Sequential Designs

Description

dsgnPrep takes Englert and Kieser's optimal adaptive design and adds information that is needed by other functions.

Usage

dsgnPrep(dsgn = NULL, w1 = "n", w2 = NULL)

Arguments

dsgn	Dataframe containing one of the designs in EKOptAdaptDesigns.
w1, w2	Stage 1 and 2 weights. If w1="n" (default), weights a calculated based on stage-wise sample sizes as described in Nhacolo and Brannath (2018). If w1="sr2", then w1=w2=1/sqrt(2).

Details

The function adds, to each x1 leading to 2nd stage, the corresponding p-value (p1) and its back-wards image (p1B), the stage-wise weights w1 and w2 and other information used in inference methods proposed by Nhacolo and Brannath (2018).

Value

Dataframe containing the input dataframe with added information.

Author(s)

Arsenio Nhacolo

References

Nhacolo, A. and Brannath, W. Interval and point estimation in adaptive Phase II trials with binary endpoint. Stat Methods Med Res, 2018.

Examples

## Not run: 
#Designs with w1a and w2 calculated based on sample sizes
EKOADwn <- data.frame()
for (j in 1:max(EKOptAdaptDesigns$id)){
  EKOADwn <- rbind(EKOADwn, dsgnPrep(dsgn = EKOptAdaptDesigns[EKOptAdaptDesigns$id==j,],w1 = "n"))
}
save(EKOADwn,file = "EKOADwn.RData")

## End(Not run)

arsenionhacolo/InferenceBEAGSD documentation built on May 9, 2019, 4:10 a.m.